Leveraging Generative Artificial Intelligence Recommendations for Image-based Chronic Kidney Disease Diagnosis

Abstract: This paper presents work leveraging the recommendations of generative artificial intelligence (AI) tools such as large language models (LLMs) to create suitable AI models for automated image-based diagnosis of chronic kidney disease (CKD) within the context of a comprehensive AI-driven healthcare system. The LLMs suggested the synthesis of image-based AI solutions such as convolutional neural networks (CNNs) and these suggestions were followed meticulously to build AI models that were then trained on computed tomography (CT) image data representing the normal kidney state as well as the presence of cysts, stones and tumors and then tasked with the diagnosis of CKD based on the classification of the input CT images. Featuring reasonable performance metrics, the resulting AI models demonstrated the effectiveness of generative AI as a tool in the synthesis, training, testing and deployment of practical AI models within healthcare settings.

Keywords: Generative Artificial Intelligence (AI), Large Language Model (LLM), Convolutional Neural Network (CNN), TensorFlow, Healthcare System, Disease Diagnosis and Prediction, Chronic Kidney Disease (CKD).